May 2025 · Agentic AI · Automation · Sales Enablement
An hour of research, in minutes.
The problem
Pre-call research takes time most sales reps don’t have. Before a discovery call, a good rep needs to know the company, the prospect, the tech stack, the competitive landscape, and the likely objections. In practice that means 30 to 60 minutes of browser tabs, LinkedIn scrolling, and manual note-taking before every single call. The reps who skip it walk in underprepared; the reps who do it properly spend hours a week on work that has nothing to do with actually selling. Either way, the business loses.
The approach
I built a sales research copilot in Relevance AI with three connected tools working in sequence. The first scrapes the company website and returns a structured summary of what the business does, how it positions itself, and what technology it appears to be running. The second takes a LinkedIn URL and builds a prospect profile covering role, background, and likely priorities. The third synthesises both into a structured pre-call report: talking points, relevant context, and suggested angles for the conversation. All sales reps do is input a company URL and a LinkedIn profile; the agent handles the rest. This was my first agent build, customised for various businesses by grounding the key prompts and context in their business goals, ICP, and offering. My marketing and sales background helped ensure the agent had the right context to work from.
The result
I pitched the idea to a founder and CTO, who piloted it for three months. Research time dropped from an hour to minutes, and reps walked into calls with accurate insights from LinkedIn and verified sources. As one rep put it: "Saved me hours pulling data together and gave me a clear picture of the tech already in play. I knew exactly how to structure my calls going in."